[Phys-L] placement / evaluation tests ... or not

[John Denker <jsd@av8n.com>]

On 12/14/2013 10:26 AM, Promod Pratap wrote:
>  I was
> wondering if there exists a test that we can administer in the first week
> of the first course of the algebra-based physics sequence that would
> predict student success in the course. 

Such tests exist, and sometimes they help, but sometimes they
cause problems.  In particular, math proficiency is a very
strong predictor of which students will do well in the course.

Sometimes there is a nasty dilemma here.  I'm not sure what is
the right thing to do.

To understand the dilemma, let's consider how this looks to
an incoming student who is not a physics major:
  They walk into a new class, and the first thing they see
  is a math quiz.  Furthermore, it is a very challenging
  math quiz, on which half the students score less than 50%.

This creates a ghastly first impression.  They assume the
rest of the course will be all the same, i.e. all math
and no physics.  They assume the course be difficult and
uninteresting.

Keep in mind that a lot of students hate math.  Even some
of the ones that like science hate math.

This is the classic pedagogical dilemma:  Something that is
helpful from a subject-matter point of view is unhelpful from
a motivational point of view.  There is a Venn diagram:
Successful teaching requires getting the subject matter right
*and* getting the motivation right.  Neither is a substitute
for the other.

I have seen placement / evaluation tests used to good effect
in situations where the students were already relatively
advanced, i.e. where they already knew a good bit of math
and a good bit of physics.  They knew how the two fit
together, and were predisposed to trust the teacher.

  In contrast, if you try the same sort of thing with a bunch
  of non-majors who are grudgingly taking the course to satisfy
  a requirement that was imposed without any credible explanation,
  it's going to make a bad situation worse.

In such a case, I don't know any easy way to solve the dilemma. 
 -- Not giving the quiz causes problems (because the teacher and
  students lack some important information).
 -- Giving the quiz causes worse problems (because it gives a
  bad first impression).

This problem would not arise if you could trust a transcript
to mean what it says, but you can't.  Students who got a
good grade in algebra class might not be able to do algebra.

When trust breaks down, lots of other things break, too.

One possibility is to start with whiz-bang activities for the
first few days (to give a good first impression) and /then/
give the quiz (to obtain needed information).  Alas, that's
far from perfect;  does anybody have a better idea????????

I see this as a cost-of-information problem.  There is high-
value information that comes at a high cost.  Ick.

> Our University will change its policy regarding drops, withdrawals, and
> academic penalties beginning next Fall. With this new policy, it would
> really help students if the could make an informed decision in the first
> one or two weeks of a course whether they would do well in it.

That exacerbates the problem.

===================

In the long run, IMHO math and physics should be taught /together/.
The physics motivates the math, and the math explains the physics.
This solves many (but not all) problems.

To make this work requires coordination between math teachers,
physics teachers, administrators, et cetera ... which is not at
all easy to set up.  This is a topic for another day.

This doesn't entirely solve the cost-of-information problem,
because you are still going to have incoming students with
widely varying backgrounds.